Version vom 23. August 2015, 15:07 Uhr

Abstract

Latest Ardumower code requires motor encoders for odometry. Using odometry you can calculate the robot's speed and the short-term position of the robot. This works by counting the rotations of left and right wheel (motor encoders - required by latest Ardumower software).

Currently, odometry is used:

• to ensure the robot left and right wheel are turning the same speed (in forward direction)

Encoder pinout (Ardumower motors)

The Ardumower has a 2-channel encoder with 5 pulses at one cycle of motor multiplied by gear ratio 1:212, resulting in 1060 encoder ticks per one full tire resolution.

Wiring/pinout of motor encoder connector:

green  ---- GND
brown  ---- VCC
blue   ---- odometry1
purple ---- odometry2

IMPORTANT: PCB v0.5/1.2 are missing pull-ups! You need to add them yourself:

odometry1 --- 4.7k --- 5v
odometry2 --- 4.7k --- 5v

Signal of odometry encoders

Forward: pin1 makes LOW->HIGH transition and pin2 = HIGH
Reverse: pin1 makes LOW->HIGH transition and pin2 = LOW

Do-It-Yourself encoders

• 

  Ambrogio L50: An encoder wheel of a computer mouse is glued to a slice. The screw of the wheel shaft is unscrewed, and the slice is added. Embedded in two Plexiglas pieces is a photo diode and photo transistor of a light barrier (e.g. LTH 301)

• 

  Tianchen TC-G158 (right): A light barrier (e.g. LTH 301) is interrupted by a gear-weheel of the motor shaft. ).

• 

  LTH301 circuit

Circuit

VCC           pinArduino
+--13 KOhm----+-------Collector  Emitter---GND
+--380 Ohm------------Anode      Kathode---GND

Differential drive

By adding (integrating) Odometry speed, we can calculate distance and heading.

• 

  Differential drive

Distance per wheel:

left_cm  = ticksLeft  / odometryTicksPerCm
right_cm = ticksRight / odometryTicksPerCm

Heading:

wheel_theta = (left_cm - right_cm) / odometryWheelBaseCm
odometryTheta += wheel_theta

Overall distance:

avg_cm  = (left_cm + right_cm) / 2.0
odometryX += avg_cm * sin(odometryTheta)
odometryY += avg_cm * cos(odometryTheta)

Ground measurement error

By adding (integrating) Odometry speed, we can calculate distance and heading. However, as we are adding a small error each time (due to not exact physical ground measurement), distance and angle errors are increasing. This can be solved by sensor fusion.

• 

  Odometry error

Motor controller (PID)

The speed of the motors is controlled by a software PID controller. You can monitor the quality of the motor speed control via pfodApp (Plot->Motor control):

Видео

Еще нет видео (функционал в разработке)